GeoMMBench 与 GeoMMAgent：面向地球科学与遥感专家级多模态智能
GeoMMBench and GeoMMAgent: Toward Expert-Level Multimodal Intelligence in Geoscience and Remote Sensing
GeoMMBench 提供了覆盖多学科、多传感器、多任务的遥感多模态问答基准，可用于系统评估 MLLM 在遥感领域的感知与推理短板；GeoMMAgent 展示了如何通过工具增强的多智能体框架（集成检索、感知、推理）提升复杂遥感任务的性能，为视觉-语言模型在专业领域的落地提供了可复用的范式。
原文依据 we introduce GeoMMBench, a comprehensive multimodal question-answering benchmark covering diverse RS disciplines, sensors, and tasks
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论近期多模态大语言模型（MLLMs）的进展加速了领域导向人工智能的发展，然而在地球科学与遥感（RS）领域，其发展仍受限于独特挑战：广泛的学科知识、异构的传感器模态以及碎片化的任务谱系。为弥合这些差距，我们提出了 GeoMMBench，一个全面的多模态问答基准，涵盖多样的遥感学科、传感器和任务，能够比先前基准进行更广泛、更严格的评估。利用 GeoMMBench，我们评估了 36 个开源和专有的大语言模型（LLMs），揭示了在领域知识、感知基础与推理——这些专家级地理空间解译所必需的能力——方面存在系统性缺陷。在评估之外，我们提出了 GeoMMAgent，一个多智能体框架，通过领域特定的遥感模型和工具，策略性地整合检索、感知与推理。大量实验结果表明，GeoMMAgent 显著优于独立的 LLMs，凸显了工具增强智能体在动态应对复杂地球科学与遥感挑战中的重要性。
英文原文摘要 1191 字符
Recent advances in multimodal large language models (MLLMs) have accelerated progress in domain-oriented AI, yet their development in geoscience and remote sensing (RS) remains constrained by distinctive challenges: wide-ranging disciplinary knowledge, heterogeneous sensor modalities, and a fragmented spectrum of tasks. To bridge these gaps, we introduce GeoMMBench, a comprehensive multimodal question-answering benchmark covering diverse RS disciplines, sensors, and tasks, enabling broader and more rigorous evaluation than prior benchmarks. Using GeoMMBench, we assess 36 open-source and proprietary large language models (LLMs), uncovering systematic deficiencies in domain knowledge, perceptual grounding, and reasoning--capabilities essential for expert-level geospatial interpretation. Beyond evaluation, we propose GeoMMAgent, a multi-agent framework that strategically integrates retrieval, perception, and reasoning through domain-specific RS models and tools. Extensive experimental results demonstrate that GeoMMAgent significantly outperforms standalone LLMs, underscoring the importance of tool-augmented agents for dynamically tackling complex geoscience and RS challenges.